Initially, dehydration of 1a to form alkene 2a was investigated. The reaction was performed in nitromethane (CH3NO₂) at 80 ℃. As shown in Table 1, in the presence of AlCl₃ catalyst, only trace amounts of the desired product 2a was detected with the starting material selectively converting to the dimerization product 3a in abundance (entry 1). Under identical conditions, FeCl3•6H2O was shown to be a less effective catalyst for the dehydration of 1a (entry 2). Toluenesulfonic acid (p-TSA) was also observed to be unreactive for this reaction (entry 3). To circumvent issues with reactivity, we studied the reaction system in the presence of an additive. It was evidenced that the addition of 5 mol% of PPh₃ in the presence of the AlCl₃ Lewis acid yielded an increase in the desired product to 93% (entry 4). However, PPh₃ cannot catalyze this transformation alone and requires the presence of a Lewis acid (entry 5). Alternative combinations of Lewis bases or Brönsted bases were also examined (entries 6-12). Tributyl phosphine (PBu3) and 1,2-bis(diphenylphosphino)ethane (DPPE) showed significantly improved efficiency compared with pyridine, o- phenanthroline and 1,4-diazabicyclo[2.2.2]octane (DABCO). To reduce cost, PPh₃ was used henceforth in our study. AlCl₃ was observed to be the most efficient Lewis acid, as when subjecting the reaction to other Lewis acids, such as p-TSA, FeCl4•6H2O or I2 in the presence of PPh₃ a dramatic drop in the conversion of 1a resulted (entries 13-15). Further investigations were also conducted as a function of the solvent. Among the various solvents tested, nitromethane clearly yielded more of the desired product, with toluene, acetonitrile and ethanol observed to be less efficient (64%-76%) (entries 16-18). Increasing the reaction temperature from 80 to 100 ℃ resulted in a decrease of the reaction selectivity, even though the reaction time was halved (entry 19). The test matrix revealed the optimal conditions to be: 5 mol% of AlCl₃ and 5 mol% of PPh₃ in nitromethane at 80 ℃ for 2 h.

Using the optimized reaction conditions, the substrate scope was probed, and the results are shown in Table 2. First, the α-phenylethanols bearing para-substituted groups were studied (entries 1-6). It was found that halide, alkyl and aryl group substituted alcohols readily yielded their corresponding olefins. The presence of an electron-donating group in the phenyl ring facilitated, to some degree, progress of the reaction (entries 1,4 and 6). However, when para-NO₂-substituted α- phenylethanol was used, only unreacted starting material was recovered. The stability of the carbocation may be ascribed to the negative effect of the electron-withdrawing group. Furthermore, ortho- and meta-substituted α-phenylethanols were also tested (entries 7 and 8), with their corresponding desired products forming with unsurprisingly good yields. It should be noted that in addition to selected terminal olefins, internal alkenes and cycloalkenes could also be synthesized using this method (entries 9-12). As shown in Scheme 1, tertiary alcohols 1b and 1c can also be selectively converted to 2n and 2o with-out the formation of any side-product. Thus the present protocol offers an efficient route to synthesize complex alkene derivatives.

However, the mechanisms by which the two catalysts cooperatively affect the catalysis remain to be delineated. TheAlCl₃-catalyzed dehydration of 1a suffered mainly from the formation of the undesirable by-product 3a. The dimerization tri-substituted alkene mechanism of formation is known to proceed in the presence of an acid catalyst through the following two pathways: (1) coupling of benzylic alcohols and styrenes, in which the styrene components act as π-type nucleophiles ^{[46, 47]}; (2) dimerization of 2a ^[48]. The two reaction pathways described to form 3a and the dehydration of 1a to 2a have been reported to be associated with the formation of a benzyl cation (I) (Scheme 2) ^{[49, 50]}. A well-established mechanism shows the benzyl cation, in the presence of the strong Lewis acid, AlCl₃, to form from either 1a or 2a; a β-H⁺ elimination of I leads to the formation of 2a; the trapping of I with 2a results in the formation of 3a. The positive effect of PPh₃ on the AlCl₃-catalyzed formation of 2a via dehydration of 1a should be linked to one or more reaction steps, in which the addition of PPh₃ results in a change to either the reaction rate or the stability of intermediate species.

To better understand the mechanism behind how PPh₃ enhances the AlCl₃ catalyst, further control experiments were performed. Initially, AlCl₃-catalyzed dimerization of 2a was examined in the presence or absence of PPh₃. The dimerization reaction of 2a was observed to be significantly inhibited in the presence of PPh₃ (Scheme 3). This result implies that, in the presence of PPh₃, the enhanced catalytic activity of AlCl₃ in promoting the dimerization of 2a was diminished, thus improving the chemical stability of 2a in the reaction system. The effect of PPh₃ on the AlCl₃ catalyst was also examined in the reaction of 1a with 1,1-diphenylethylene 4a, which is known to be an active π-nucleophile, however, inert toward the dimerization reaction ^[51]. With AlCl₃ alone, the Friedel-Crafts-type alkylation product 3b was isolated in 62% yield along with the formation of 22% of 3a. There was no discernible detection of 2a in this system. In the presence of PPh₃, the formation of 3a was totally suppressed, and additionally, the yield of 3b decreased to 36%. However,2a could be isolated in 41% yield. These results implied that the addition of PPh₃ had a detrimental effect on the Friedel-Crafts-type reaction of 1a and 4a. However, the formation reaction of 2a benefited from the use of PPh₃. Further control experiments were performed including the AlCl₃-catalyzed coupling reaction of benzhydrol 1d with 4a. As the benzyl cation generated from 1d is chemically inert toward -H+ elimination, the results were expected to directly reflect the effect of PPh₃ on the AlCl₃-catalyzed Friedel-Crafts alkylation of 4a. It is not a surprise to observe that AlCl₃ is an excellent catalyst for promoting such a reaction, and the expected tri-substituted alkene 3c was formed in 98% yield after 2 h of reaction. However, after the addition of PPh₃, the yield of 3c significantly decreased to 65%. Considering unreacted 4a can be recovered, we therefore concluded that the PPh₃ Lewis base was effective in decreasing the rate of the Friedel- Crafts-type nucleophilic substitution reaction. Finally, dehydration of 1-phenylpropanol 1e was investigated. The poor nucleophilicity of the generated product,1-phenyl-1-propene 2j, resulted in suppression of its corresponding dimer being formed. Therefore, this reaction can directly reflect the effect of PPh₃ on the AlCl₃-catalyzed dehydration of 2-arylethanol. Furthermore, it was evidenced that the yield of 2j decreased from 97% to 93% in the presence of PPh₃ in the reaction system. This result indicated that the addition of PPh₃ has no significant effect on benzyl cation intermediate formation and the subsequent β-H⁺ elimination step.

Collectively, all the results detailed provided an improved understanding as to circumvent the promoting effect of PPh₃ on the AlCl₃-catalyzed dehydration of 1a. In the absence of PPh₃, the reaction rate to form 3a is faster than H+ elimination. Therefore,3a is the major product. The presence of PPh₃ likely decreases both the rates of nucleophilic substitution and β-H⁺ elimination. However, the likelihood is for a more significant drop in the rate of the former reaction. Therefore, the H+ elimination reaction is relatively faster than 3a formation (Scheme 2) giving rise to a significant change of product distribution—hence the dramatic improvement of selectivity to 2a.

The carbocation intermediate can be stabilized by polar and aprotic solvents, such as nitromethane, acetonitrile and dichloromethane ^[52-55]. The intrinsic electrostatic interaction between the carbocation and the negative center of the solvent molecule is responsible for this stabilization effect. In the presence of PPh₃, the generated benzyl cation (I) may interact preferentially with PPh₃ because of the existence of a lone pair of electrons (Scheme 4). Such interaction types not only ensures improved stability of I, but additionally blocks access of the nucleophile to I, which slows the rate of the Friedel-Crafts-type nucleophilic substitution reaction. As β-H⁺ elimination follows an E1 mechanism (single molecular reaction), it is not significantly affected by this type of interaction.

Reviewing the results in hand led to consideration of downstream applications of this Lewis base-assisted Lewis acid-catalyzed alkene formation via dehydration of alcohols. 2-Cinnamyl-1,3-dicarbonyl compounds are important synthons for organic synthesis ^{[56, 57]}. The synthesis of such compounds is therefore a very active research topic. The reported methods hitherto can be categorized into the following four routes: (1) nucleophilic substitution of cinnamyl halides with 1,3- dicarbonyl compounds under basic conditions ^{[58, 59]}; (2) acid- catalyzed direct C2-allylation of 1,3-dicarbonyl compounds with cinnamyl alcohols or their esters ^{[60, 61]}; (3) Pd-catalyzed Trost-Tsuji reactions of cinnamyl carbonate or acetates ^[62]; similar reactions have also been reported using Pt/pyrrolidine as the catalyst ^[63]; and (4) direct allylic alkylation of 1,3-dicarbonyl compounds with alkenes via Pd(II)-catalyzed allylic C-H activation ^[64]. Although various routes to synthe-size 2-cinnamyl-1,3-dicarbonyl compounds have been reported, those methods suffer from limitations of substrate generality, the availability of starting materials, multistep synthesis, use of expensive catalysts, the requirement of ligands or additives, low product yields, and the need for harsh synthesis conditions. Therefore, the design of improved and efficient approaches that allow for the rapid, cost-effective synthesis of 2-cinnamyl-1,3-dicarbonyl compounds from readily available precursors is highly desired.

Conversely,2-aryl-3,4-dihydropyrans are known to be easily synthesized through a three-component reaction of styrene,1,3-dicarbonyl compounds and formaldehyde in water in the absence of any catalyst ^{[65, 66]}. Such widely accessible compounds led us to explore the possibility of their use in organic synthesis as building blocks ^[67-70]. Unexpectedly,2-phenyl- 3,4-dihydropyran 5a can be converted to 2-cinnamyl-1,3- dicarbonyl compound 6a in the presence of the mild Lewis acid, MnCl2. Unfortunately, the yield is minimal (Scheme 5). This transformation was theoretically triggered by the acid-catalyzed cleavage of an endocyclic C-O bond of the 3,4- dihydropyran, which led to the formation of a benzylic cation (II). The following β-H⁺ elimination of the carbocation intermediate resulted in the formation of 6a. We speculate that the potential nucleophilicity of the 1,3-dicarbonyl fragment is partially responsible for the low yield of this reaction, which enabled intramolecular and intermolecular trapping of the benzylic cation to be possible, thus decreasing the reaction selectivity of 6a formation.

The strategy of the above-mentioned Lewis base-assisted Lewis acid-catalyzed alkene formation via alcohol dehydration was taken and applied to improve the yield of this reaction. However, the AlCl₃/PPh₃ system was observed to be inappropriate for this reaction to proceed efficiently (Table 3, entry 1). Further screening the reaction parameters found that the common Lewis acid, I₂, in the presence of PPh₃ resulted in significant conversion of 5a to 6a yielding 88% in CH3NO₂ (entry 2). In the absence of PPh₃, only 23% yield was obtained (entry 3). Multiple organic bases were examined, with PPh₃ proving to be the most efficient (entries 4-8). The I₂/PPh₃ reaction system could be scaled up to multigram quantities yielding similar conversions (entry 9). This system can be further extended to convert various other dihydropyran derivatives (Scheme 6). Given the large number of commercially available 1,3- dicarbonyl compounds, which act as starting reagents, together with the dihydropyran derivative ease of preparation, this synergistic protocol was expected to be a practically useful method to synthesize 2-cinnamyl-1,3-dicarbonyl compounds.